Dev uses AI-tuned document camera to identify tabletop miniatures for YOLO training
TheZachMueller · x · 2026-11-12
Developer TheZachMueller shares test one of his AI-augmented Descent 2e setup: identifying tabletop miniatures from top-down profiles via a document camera.
- Having his AI assistant (Sol) tune the camera worked "shockingly well": focus locked at 190, sharpness 6, white balance fixed at 4600 K to kill the orange cast
- The tuned 8 MP stream plus raw stream/stills are exposed for model training
- Next step: scale from a 16" touchscreen to a 43" IR TV and collect data to train a YOLO model
Related event: Developer Augments Descent 2e Solo Play with AI and Computer Vision(2 posts)→
More from coding & agent
- Together Compiler: agent-driven continual compilation for AI inference — togethercompute · 2026-11-12
- Grok agent autonomously browses, orders and pays for a book on an unfamiliar site — xkonjin · 2026-11-12
- Stop Managing Sandbox Fleets for Cloud Agents, Argues gakonst: Go Harness-First — andreisavu · 2026-11-12
- Claude Opus 5.5 hunted down 360 bugs in Mediabunny, catching every user-reported one early — Vjeux · 2026-11-12
- Mediabunny bug hunt nears end: 354 accepted changes, huge robustness gains — Vjeux · 2026-11-12
- Sentry CEO: Opus 5.5 Is a 'Step Function Change' for Coding — Switch If You Can Afford It — zeeg · 2026-11-12